Historical trends in the risk of spring frost damages to fruit trees in Eastern Canada

As one of the consequences of the global warming, growing season has become longer with an earlier start for both annual and over-wintering crops in Canada. An earlier start of growing season implies that the phenological development of crops could have been advancing. It has been suspected that crops, especially the over-wintering perennial crops such as fruit trees, could become more vulnerable to spring frost damages. More speculations were stimulated by extreme events in recent years, such as spring frost damages to fruit trees in 2007 and 2012 in the United States and Eastern Canada, together with a potential increase of climate extremes in the projected future climate. However, historical trends in the risk of spring frost damages were seldom reported, given the rarity of major frost events and very limited phenological observations. In this study, phenological stages of apple trees, a major fruit crop in Eastern Canada, were estimated from historical daily temperature data through phenological models recently developed with laboratorial and field experiments. Using a homogenized historical daily temperature dataset, dates when the apple crops met their chilling requirement and forcing requirements to reach their phenological development stages of bud break (green tip), ½ inch green, tight cluster, first pink, full pink, first bloom and full bloom, were estimated for each year, respectively. A non-parametric trend analysis method was employed to estimate historical trends in these dates for the development stages of apple crops and the last spring occurrences of two critical low temperatures (-2.2°C and -3.9°C) that are corresponding to 10% and 90% kills to the buds at the later flowering stages after full pink. Results showed that all the dates became 8-10 days earlier throughout the past 110 years for all phenological stages from bud break to full bloom although a significant trend was not observed in the dates when the chilling requirement was met as widely as other phenological stages. Meanwhile, the last spring occurrences of the two critical low temperatures were also observed by approximately 11 and 15 days earlier. The risk of frost damages was estimated by the probability of a critical low temperature occurring at a phenological stage that could result in a 10% or 90% kill to the buds. Both significant positive and negative trends were found in the risk of a 10% kill at phenological stages ½ inch green, first pink and full pink but no significant trends were found in the risk of a 90% kill. Therefore, it was difficult to draw a conclusion whether the risk of spring frost damages became higher or lower to apple crops in Eastern Canada.